Thursday, 19 May 2016

The referendum polls are even worse for Leave than they look

Those EU referendum opinion polls: today the economic analysts at Lombard Street Research examine the data in a daily note. 

I can’t give you a link, but here is the headline I’d put it: ‘The undecideds mean Brexit wipe-out.’

LSR makes it more wordy, of course, but in the end they get to the same place: ‘Everyone likes a close race, and the media are trying really hard to portray the upcoming referendum on the UK’s membership of the EU as one that could go either way. Polls are often used as evidence of this.’

‘According to the FT, the average percentage of Leave responses in surveys carried out since 2010 is 40%, vs 42% for Remain. More recent polls since last September show a clearer Remain advantage – 44.0% vs 40.5% – and data collected by Bloomberg over the same period show a similar gap – 43% vs 39.3%.’

But don’t just look at those Leave/Remain preferences. There’s more to analysing the figures than that:

‘…this bigger lead for the Remain camp is fairly small compared to the standard deviation on the responses (four to five percentage points), making it impossible to point to a clear winner with any degree of confidence.’

LSR clearly wrote this note before the publication yesterday of the Ipsos MORI poll, carried out over the weekend. Not that Ipsos MORI makes it look any better for Brexit. The poll showed the Remain campaign with an 18-point lead over Leave.

A fluke poll? No. The figures could in the end be every bit as bad as that, because, as LSR continues: ‘Further analysis of the data, however, provides more insights. In most surveys the percentage of Undecided votes is very high – typically between 15% and 20%.’

‘With both Remain and Leave well below 50%, it is clear that it’ll be the people who haven’t made up their minds yet who will determine the outcome. And it’s the analysis of the Undecided voters’ inclinations and preferences that provides the most compelling clues that a Brexit is very unlikely.’

This is because, according to further data from a Bloomberg tracker poll, when more people have expressed a clear Remain or Leave preference, there has tended to be a higher percentage of those in favour of the UK remaining part of the EU.

A similar pattern is evident in the data collected by the Financial Times, though the pattern was not as clear as in the Bloomberg data:the data show most undecideds go Remain.

In other words, as in most referendums, the status quo wins – more evidence that the Leave campaign has not yet got it through the heads of the voters that Remain does not mean a UK-EU status quo. But that is another issue.

Of course, we learned in the general election not to rely on opinion poll data, because the figures can be remarkably inaccurate. But other data supports the LSR conclusions.

The Number Cruncher Politics website admits the volatility in the polls, but then says that increased support for remain is down to Tory voters swinging back to Remain:

‘The tables from the MORI poll shows a huge swing about Conservative voters compared with last month, when Leave led 50-42 among those that would vote Conservative in new election now. This month, Remain leads 60-34 among Tory voters – a greater margin than the overall average.’

In short, wipe-out for Brexit.

1 comment:

  1. Sound logic to me. The status quo or inertia will be to Remain and so Vote Leave will have lost us a winnable referendum by making every single mistake it is possible to make.